From d71d7527a8be8654560867d761036598408bfe14 Mon Sep 17 00:00:00 2001
From: Robin Haberkorn <rhaberkorn@fmsbw.de>
Date: Mon, 29 Jun 2026 19:14:03 +0200
Subject: implemented tere_set_is_interrupted_cb()

Allows you to set up a callback that's invoked repeatedly during
matching. If it returns non-0 matching will abort and REG_EINTR is
returned.
Allows you to interrupt long-running regexp executions that would
otherwise block the UI or hang the entire application.
---
 README.md  |  8 +++-----
 regcomp.c  | 11 +++++++++++
 rege_dfa.c | 12 ++++++++++++
 regerrs.h  |  1 +
 regex.h    |  2 ++
 regexec.c  |  5 +++++
 regguts.h  |  5 +++++
 7 files changed, 39 insertions(+), 5 deletions(-)

diff --git a/README.md b/README.md
index c35480e..6dbfbca 100644
--- a/README.md
+++ b/README.md
@@ -23,14 +23,12 @@ Compared to hsrex, this library has the following changes:
   `tere_comp()` compilation flag.
 * Support the `REG_ANCHORED` flag for `tere_exec()`.
 * `tere_free()` ignores nullified `regex_t` objects.
+* `tere_set_is_interrupted_cb()` allows configuring a repeatedly
+  invoked callback. If it returns true (non-null) matching aborts
+  immediately and `tere_exec()` will return `REG_EINTR`.
 
 ## TODO
 
-* Hook into the matching algorithm.
-  Even though runtime cannot be as catastrophically bad as in pure
-  backtracking engines like PCRE, `tere_exec()` calls can still be slow
-  on extremely large texts.
-  The hook allows interruptions.
 * Expose enough API to swap out the regular expression lexer.
   Useful for custom DSLs like glob patterns or TECO patterns.
 * Support splitting the subject string into two halves, so we can
diff --git a/regcomp.c b/regcomp.c
index bd332dc..e6c9f94 100644
--- a/regcomp.c
+++ b/regcomp.c
@@ -272,6 +272,17 @@ static struct fns functions = {
     rfree,			/* regfree insides */
 };
 
+/*
+ - tere_set_is_interrupted_cb - set is_interrupted callback
+ ^ void tere_set_is_interrupted_cb(regex_t *, int (*cb)(void));
+ */
+void
+tere_set_is_interrupted_cb(
+    regex_t *re, int (*cb)(void))
+{
+    ((struct fns *)re->re_fns)->is_interrupted = cb;
+}
+
 /*
  - compile - compile regular expression
  ^ int compile(regex_t *, const chr *, size_t, int);
diff --git a/rege_dfa.c b/rege_dfa.c
index a055921..8c6f2fc 100644
--- a/rege_dfa.c
+++ b/rege_dfa.c
@@ -101,6 +101,9 @@ longest(
 	    if (ss->flags & NOPROGRESS && v->eflags & REG_ANCHORED) {
 		return NULL;
 	    }
+	    if (IS_INTERRUPTED(v->re)) {
+		return NULL;
+	    }
 	    cp = nextchr(v, cp);
 	    ss->lastseen = cp;
 	    css = ss;
@@ -119,6 +122,9 @@ longest(
 	    if (ss->flags & NOPROGRESS && v->eflags & REG_ANCHORED) {
 		return NULL;
 	    }
+	    if (IS_INTERRUPTED(v->re)) {
+		return NULL;
+	    }
 	    cp = nextchr(v, cp);
 	    ss->lastseen = cp;
 	    css = ss;
@@ -240,6 +246,9 @@ shortest(
 	    if (ss->flags & NOPROGRESS && v->eflags & REG_ANCHORED) {
 		return NULL;
 	    }
+	    if (IS_INTERRUPTED(v->re)) {
+		return NULL;
+	    }
 	    cp = nextchr(v, cp);
 	    ss->lastseen = cp;
 	    css = ss;
@@ -261,6 +270,9 @@ shortest(
 	    if (ss->flags & NOPROGRESS && v->eflags & REG_ANCHORED) {
 		return NULL;
 	    }
+	    if (IS_INTERRUPTED(v->re)) {
+		return NULL;
+	    }
 	    cp = nextchr(v, cp);
 	    ss->lastseen = cp;
 	    css = ss;
diff --git a/regerrs.h b/regerrs.h
index 259c0cb..6d23e42 100644
--- a/regerrs.h
+++ b/regerrs.h
@@ -17,3 +17,4 @@
 { REG_MIXED,	"REG_MIXED",	"character widths of regex and string differ" },
 { REG_BADOPT,	"REG_BADOPT",	"invalid embedded option" },
 { REG_ETOOBIG,	"REG_ETOOBIG",	"nfa has too many states" },
+{ REG_EINTR,	"REG_EINTR",	"interrupted" },
diff --git a/regex.h b/regex.h
index 5568ed7..d48f20e 100644
--- a/regex.h
+++ b/regex.h
@@ -290,6 +290,7 @@ typedef struct {
 #define	REG_MIXED	17	/* character widths of regex and string differ */
 #define	REG_BADOPT	18	/* invalid embedded option */
 #define	REG_ETOOBIG	19	/* nfa has too many states */
+#define	REG_EINTR	20	/* interrupted by is_interrupted() callback */
 /* two specials for debugging and testing */
 #define	REG_ATOI	101	/* convert error-code name to number */
 #define	REG_ITOA	102	/* convert error-code number to name */
@@ -320,6 +321,7 @@ MODULE_SCOPE int __REG_WIDE_EXEC(regex_t *, __REG_CONST __REG_WIDE_T *, size_t,
 #endif
 MODULE_SCOPE re_void regfree(regex_t *);
 MODULE_SCOPE size_t regerror(int, __REG_CONST regex_t *, char *, size_t);
+MODULE_SCOPE void tere_set_is_interrupted_cb(regex_t *, int (*cb)(void));
 /* automatically gathered by fwd; do not hand-edit */
 /* =====^!^===== end forwards =====^!^===== */
 
diff --git a/regexec.c b/regexec.c
index aadb382..ae6ca77 100644
--- a/regexec.c
+++ b/regexec.c
@@ -328,6 +328,11 @@ exec(
 	memcpy(VS(pmatch), VS(v->pmatch), n*sizeof(regmatch_t));
     }
 
+    if (IS_INTERRUPTED(re)) {
+	/* normalize return code in case of interruptions */
+	st = REG_EINTR;
+    }
+
     /*
      * Clean up.
      */
diff --git a/regguts.h b/regguts.h
index 0fd46c4..e941c28 100644
--- a/regguts.h
+++ b/regguts.h
@@ -385,8 +385,13 @@ struct subre {
 
 struct fns {
     VOID FUNCPTR(free, (regex_t *));
+    int FUNCPTR(is_interrupted, (void));
 };
 
+#define IS_INTERRUPTED(re) \
+	(((struct fns *)(re)->re_fns)->is_interrupted && \
+	 (*((struct fns *)(re)->re_fns)->is_interrupted)())
+
 /*
  * the insides of a regex_t, hidden behind a void *
  */
-- 
cgit v1.2.3